The Digital World Is Not a Flat Circle
TheFamily Papers #001
By (Co-Founder & Partner) |
Back in 2007, Pankaj Ghemawat made waves by pushing back against the central point of Thomas Friedman’s bestseller, The World Is Flat. As Ghemawat brilliantly put it in an article, then a book, then a subsequent TED talk, the world is not flat, but rather it is uneven. Whether you’re a business or an individual, crossing borders creates so much friction that the world economy is not global so much as it is local (or regional for that matter). This is the reason why building a successful worldwide business is not a universal recipe. Expanding a company’s operations demands constant adaptation to the specific characteristics of various domestic markets.
Why the digital world appears flat
In sharp contrast with Ghemawat’s argument, people in the digital economy continue to see the world as flat. The era of Facebook and Google is marked by a lack of borders and friction, global businesses serving the global Internet. Consequently, every startup founder has probably heard this at some point: “Your startup,” they say, “should be global from day one.”
That sounds all well and good, but there’s a hitch: building global operations is becoming more and more difficult. The idea of a flat digital world is an illusion, and the culprits are American pioneers. They scaled up the first global tech companies, setting an inspiring, yet misleading, precedent. The global startup fallacy relies on five contingent facts:
- in the nascent digital economy, Silicon Valley had an early start due to the legacy of the semiconductor and computer industries and the availability of critical resources (know-how, capital, and rebellion). Those early Entrepreneurs were the first to scale up, never encountering serious competition for foreign markets;
- as they went global, those pioneers benefited from 60 years of US investment in soft power, from the Marshall Plan onwards. As we all watched American movies and listened to American music, it seemed all the more natural to adopt American startups’ early value proposals. Like those who bought American jazz records in the 1960s, a lot of people all around the world were willing to adopt new Internet applications simply because it was something new coming from the US;
- as pointed out by Paul Graham, the cost of founding a software startup has fallen to zero, which frees up capital to finance the best startups’ global expansion once they have found their product-market fit. Similarly, thanks to spectacular progress in technology, most friction (marketing, distribution) has been reduced to a minimum for digital startups;
- as friction disappeared, so did barriers to entry: the result was that the large scale of operations and accelerating growth rate have become critical to ensuring domination. Since digital startups are so vulnerable, they have no choice but to be global. They should operate a global value chain, carefully choose their centers of operation, and limit their local presence to a minimum. A few companies mastering this field gave the impression that it was easy;
- finally, that first generation tackled industries that were both intangible and unregulated. The first industries to be eaten by software, such as advertising, helped Silicon Valley warm up before tackling more difficult challenges, such as entering the healthcare market or disrupting the car industry.
Thanks to this early start, the US has tied up an impressive amount of knowledge dedicated to helping founders project their operations onto a global scale. For instance, a remarkable collection of advice was put together by venture capitalist John O’Farrell a few years ago. Other countries have applied the recipe, with Israel being the main example: mastery of the English language + privileged relationships with the US through the Jewish diaspora + attracting US venture capitalists with the Yozma program = Israel securing itself onto the tail of the US-based global digital economy.
Bumps in the road
Just as going global has become more of an imperative, many signals make it clear that it is also getting more and more difficult. Here are five of those signals, all of which push forward the idea that the digital world may not be so flat after all:
- Fab exploded in 2013 while trying to expand its operations in Europe. It was an early example of a unicorn failing in its quest for a dominant global position;
- Alibaba crushed eBay in China, halting the growth of a tech giant that had previously appeared invincible. Because they were Chinese entrepreneurs trying to solve problems for Chinese customers, the Alibaba people created a for more appealing experience for the Chinese market than that designed by Americans clumsily trying to replicate their own recipes on foreign markets. The movie The Crocodile in the Yangtze tells the story in great detail;
- Homejoy closed down last summer. Many explanations, some contradictory, were given as to the reason why, but one major problem was due to the promising startup coming up short in its epic quest to open up operations in dozens of cities over three different countries;
- Uber faces so many different sets of local regulations (and so many lawsuits) that it constantly appears to be on the verge of falling down. As it nevertheless keeps on growing and taking more risks, we are probably witnessing the riskiest late-stage investment rounds in the history of the post-bubble digital economy;
- the very existence of Rocket Internet testifies as to how difficult it is becoming to go global. If it were as easy to scale up as it once was for Yahoo or Google, then there would be no room for the Samwer brothers to launch clones of successful startups on underserved markets.
In retrospect, two things explain why the digital world suddenly appears less flat: startups are operating more tangible businesses; and they are entering more regulated markets.
The tangible digital world
Every digital startup has a tangible side. At the very least, it employs engineers, developers, and designers; and if it serves hundreds of millions of users, like Google or Facebook, it deploys data centers and content delivery networks. Producing content is yet another step in the direction of tangibility. Early content producers, such as Yahoo, have historically been topped by pure middlemen such as Google and now Facebook, precisely because producing content damaged the model’s scalability.
Many tangible startups (Kozmo is a famous example that comes to mind) had a hard time taking off before the age of the smartphone and geolocation. The first successful tangible startup was Amazon, with its large infrastructure of warehouses and logistics platforms. Because of those tangible assets, Amazon is less scalable than its intangible counterparts — for a long time, it wasn’t even considered to be a tech company. The good news for Amazon is that tangible assets serve as a barrier to entry. The bad news, as demonstrated by Fab, is that a heavy, rigid, tangible infrastructure can sink the entire business.
To prevent such a fate, tangible digital businesses have to compensate by generating superior increasing returns on the digital side of their business model — the key being, as Tim O’Reilly pointed out, an exceptional customer experience that invites user participation. Consumer businesses, even though they’re subject to a very high pressure on prices and margins, perform better on that front: there are more powerful network effects on consumer markets. Conversely, enterprise markets may be a trap: FedEx sells to companies because it’s expensive (no ordinary individual can pay for it on a regular basis); and it’s expensive because its scalability is limited by the absence of user participation on an enterprise market.
Tangible startups live in a world that isn’t flat at all. They have to locate their tangible assets somewhere, and scarcity of resources demands that their deployment should be concentrated on a narrow area. They also have to enroll users, which demands marketing the product in the local language and making a considerable investment in content.
More and more digital startups are entering the tangible zone, because tangible assets are a means to an end: scaling up the product. Munchery is a promising case in point, with their “internal mission of making real, good food accessible to everyone, everywhere.” A key Munchery feature is proximity, which makes it difficult to operate solely as a global value chain. Apple is another example: why did a company with such a powerful global brand even bother opening stores, renting real estate, hiring hundreds of people under local labor laws? Because, as Babak Nivi wrote, “the best products require unique means of scaling. The delivery of the best products is tied into the product itself. Apple’s efforts to develop new manufacturing techniques and stores for its products” have but one explanation: the will to scale the product even more.
The regulated digital world
Then there are the regulations, which create bumps in the road even for intangible startups. Netflix’s operations are mostly intangible: they are deployed on Amazon Web Services, thus leaving the tangible part to Amazon. As such, it would seem that they could restrict themselves to a pure global value chain (imagine a global Netflix on which you could watch everything without a VPN). Unfortunately, they have to comply with local regulations that force them to tweak their value proposal from country to country and to hire local teams, in no small part for marketing, PR, and negotiating rights. The world is not flat if it is crumpled up by domestic regulations.
Permission is probably the trickier question in the regulation field. Permission is about waiting for the government to deliver its authorization to operate. Whenever a startup asks anything from the government, it tends to get messy. Permission can be refused for the wrong reasons, such as corporatism or even corruption. And in any case, every business takes a hit when confronted with delays and bureaucracy.
In a world where government officials thought like Nick Grossman, of Union Square Ventures, regulation by permission would be replaced by data-driven accountability (ongoing monitoring of real-time data), thus enabling innovative business models to take off while enforcing oversight by the government (see detailed proposal here). But as long as we’re not using the Internet way to regulate things, the best practice appears to be permissionless innovation: for any entrepreneurs searching for a new business model, it’s probably better to move forward, get ahead and settle things afterwards — as Airbnb did with the French hotel tax (as well as in others areas, including its home base of San Francisco).
Compliance is less of a constraint than permission. The idea is not to wait for the government approval before you can even launch operations, but to start a business and to make sure that rules are enforced at every stage. If a startup is beyond product-market fit, it shouldn’t be a problem to raise money and cover the cost of compliance. Conversely, refusing to comply would probably impede the efforts to scale up.
Fortunately, compliance is getting more and more commoditized. As Marc Andreessen and Ben Horowitz suggested here, artificial intelligence is so powerful now that software could swallow any regulation and implement compliance right into any application, making it easy to comply with local regulations almost in real-time and for a minimal cost. So instead of changing regulations and waiting for the states to harmonize them in the entire world, it’s actually easier to develop a proxy that would allow any startup to comply in exchange for almost nothing. Full stack startups in highly regulated sectors such as healthcare or education are currently laying the groundwork.
In other words, compliance-as-a-service is on the way and will help regulated startups to finally be global from day one. This has already been explored with copyright clearing platforms (the Getty Images API being a good example) and progress has been made on the legal and tax compliance front for the self-employed (seen with Intuit). But it could concern any regulation. Jeff Bezos joked in 2006 about turning tax lawyers into a web service. Tim O’Reilly alluded to the commoditization of tax compliance in this old blog post about the so-called “Amazon tax”:
In an imaginary world where Jeff Bezos was as public spirited as he is far-sighted about pursuing competitive advantage, Amazon would not only willingly collect and pay sales tax, but would offer the infrastructure they built for doing so to other online businesses.
Taxes are actually a whole category by themselves. Submitting to domestic tax rules really depends on a company’s status regarding the tangible / intangible. An intangible startup should make a priority out of maximizing its returns while operating a global value chain with minimum tax exposure in the various countries. On the contrary, a startup operating a tangible business and/or on a regulated market should think twice before making a priority out of over-planning its taxes. In some cases, it’s better not to try to have it both ways, but rather to sacrifice tax planning for the sake of scale.
Amazon felt the pain when its self-imposed constraint of not collecting sales taxes in the US prevented it from opening new warehouses in the most populated states. It ultimately conceded defeat and turned it into a victory: Amazon agreed to collect the taxes, then went on to open many more warehouses and pursue the goal of expanding same-day delivery. Contrary to popular wisdom, paying taxes is not necessarily bad for business: quite the contrary, if you’re the leader on the market, you can compensate ten-fold by increasing your returns even more and widening the gap with your closest challengers. (More recently, Amazon has announced that it will now assess its profits separately in each EU member state instead of funneling all its revenue through Luxembourg.)
Capital
Scaling up a global business takes a lot of money. The main reason why European startups fail to achieve that goal (BlaBlaCar being one of the rare exceptions) in an uneven digital world is not the size of their domestic market, but a lack of capital to fund their growth and keep on taking bold risks beyond product-market fit. In the US, on the other hand, capital is not a problem. Return on invested capital put towards conquering a dominant global position is almost infinite if returns keep on increasing at scale: this is why Uber has raised more than $5B since founding in 2009.
Capital not being a problem does not mean that investors should waste their money on anything. But financing a startup aiming at global domination looks more and more like uncharted territory, to the point that we’re seeing so-called global companies effectively operating joint ventures together with local investors to conquer the largest domestic markets (see Uber and Airbnb in China). As William Janeway wrote earlier this year,
New services are delivered physically and locally, unlike the virtual services delivered by Google and Facebook. Consequently, as each has been learning city by city, they are subject to local ecosystems that have evolved around the provision of such services: regulatory, political, cultural. The London taxi industry is not the same as that of New York nor San Francisco. The unbounded growth rates implicit in their valuations are subject to exogenous impediments that may induce step-function revaluations of growth prospects.
A powerful ally: the multitude
Whether tangible or regulated, every business has a powerful ally when trying to maximize its return on invested capital: its user community. Enrolling individuals on a platform allows the company to transfer the investment effort on the users themselves (we pay for our cars, not BlaBlaCar) while also making it easier to circumvent regulations or at the very least to maintain individual activity below certain legal thresholds. It also helps to cope with uncertainty while conducting more rigid operations. As proved by Uber and Airbnb, “sharing economy” platforms are probably the only viable business models aimed at global scale of operations in highly tangible and regulated industries. (But beware the reverse network effects.)
Why go global?
Why is it even important to go global? We long lived in a world in which brands and businesses were different from one country to another. For one McDonald’s or Coca-Cola, both famous global brands, there were many companies, such as the mighty Walmart, which operated in only a few countries.
It’s different in the digital economy — and increasing returns are the culprit. In the digital world, flat or not, the more you grow, the faster you grow — as opposed to what happens in the traditional business world where the more you grow, the more you run out of steam. When a company is driven by increasing returns, it has no choice but to reach a global scale. Expanding globally not only accelerates growth, it also ensures the highest returns at scale. Conversely, remaining on certain domestic markets only raises the probability of a competitor gaining more velocity elsewhere and ultimately crushing slower-growing companies with superior returns at scale.
Going global shouldn’t be a compromise — a startup can’t be different companies depending on the country. If it were different in, say, France and Germany, it would mean there would be almost no cross-border network effects. Even though a startup will encounter different customers on distant markets, its product and its culture should be the same across borders, if only to increase returns even more.
Again, in the end, building a successful global business in an uneven world is not a universal recipe, it is rather a formula that will likely be reinvented in each case. This is a challenge for ambitious and creative Entrepreneurs: surely those have a bright future before them!
(This is an issue of TheFamily Papers series, which is published in English on a regular basis. It covers various areas such as entrepreneurship, strategy, finance, and policy, and is authored by TheFamily’s partners as well as occasional guest writers. Thanks to Oussama Ammar, Kyle Hall, Balthazar de Lavergne, Laetitia Vitaud, and Tyler Willis for reading drafts of this and contributing suggestions.)
